A Chilled-Water-Cooler-Capacity Model for Data Center CFD and Other Cooling Digital Twins (ASHRAE 2025 Annual Conference, PX-25-075)

Overview

Paper PX-25-075 by James VanGilder, Michael Condor, Wei Tian and Hossam Mohamed Alaa Barima was presented at the ASHRAE 2025 Annual Conference (Phoenix) and appears in ASHRAE Transactions Vol. 131 Pt. 2. This record is metadata-only and is based on ASHRAE's published discussion record; the full paper is paid. The paper addresses how to predict the capacity of chilled-water air-cooling units at any operating point for data centre CFD and digital twins.

The paper presents a physics-based model that captures how cooling-coil effectiveness varies with both air and water flow, so that a single rated capacity point can be extrapolated to other water flows and temperatures. According to the authors' response, the difficult part is modelling coil effectiveness as a function of both air and water flow rates. At fixed water flow, capacity scales linearly with return-air minus entering-water temperature (neglecting fan heat) and roughly as a power of airflow.

The model lets a user enter one rated capacity at one operating condition and predicts capacity at any other operating point, including different chilled-water flow rates. The authors contrast this with simpler semi-empirical approaches that assume fixed water flow or interpolate between rated points. In the discussion, a commenter noted that at least one commercial DC CFD tool already models variable chilled-water flow effects; the authors distinguished their rigorous physics approach from interpolation.

The limit of this record is that only the discussion record, not the paper itself, has been summarised, so model equations and validation results cannot be relayed here. The related standard is ASHRAE Standard 127. The practical takeaway is that in CFD and digital twin work a cooler's rating should be treated not as a fixed number but as a capacity that depends on operating conditions.

Key points

1. The model lets a user enter one rated capacity at one operating condition and predicts capacity at any other operating point, including different chilled-water flow rates.
2. The difficult part is modelling coil effectiveness as a function of both air and water flow rates.
3. At fixed water flow, capacity scales linearly with return-air minus entering-water temperature (neglecting fan heat) and roughly as a power of airflow.
4. A commenter noted that at least one commercial DC CFD tool already models variable chilled-water flow effects; the authors distinguished their physics approach from interpolation.
5. Presented at the ASHRAE 2025 Annual Conference (PX-25-075, ASHRAE Transactions Vol. 131 Pt. 2).
6. The record is metadata-only, based on the published discussion record; the full paper is paid.
7. Related standard: ASHRAE Standard 127.

Using it in design

• Chilled-water CRAH/fan-wall catalogue items should store rated capacity with its rating conditions (EWT, return air temperature, airflow, water flow).
• That way the tool can derate capacity at the project's actual water temperature and flow rather than using the nameplate figure.
• Define water flow as a variable parameter in CFD and digital twin inputs.
• Show the user a warning where the fixed-water-flow assumption does not hold.

Who should read it

Mechanical engineers and tool developers who build or use data centre CFD and cooling digital twins.

Related standards

• ASHRAE Standard 127

Publication details

Authors

James W. VanGilder; Michael Condor; Wei Tian; Hossam Mohamed Alaa Barima

Publisher

ASHRAE (ASHRAE Transactions Vol. 131 Pt. 2)

Year

2025

Document number

PX-25-075

Event

ASHRAE 2025 Annual Conference

Document language

English

Source type

Scientific

Access

Citation only

Licence

Copyright retained; citation only — the publisher’s terms apply to the original document

Citation